GA4 alternatives for small sites: choose by question, not features

Short answer: GA4 is not the wrong tool — it is the wrong size for most small sites. If you only need to know where your traffic comes from, which pages earn their keep and where visitors get stuck, a lighter tool answers faster. Keep GA4 when you buy ads, build audiences, or have someone paid to analyse data.

Do you actually need GA4?

GA4 is a strong product and the standard free option. It is also built for a marketing department: audiences for remarketing, attribution models, explorations, a raw event export into BigQuery. If someone's job is to run it, that depth is the point — this is not an argument against it.

The mismatch appears on a small site, in three places.

• Setup and upkeep. A tag, a consent mode, key events to name and maintain, a reporting interface to learn. Every one of those is a real task on a site run by one or two people.
• Privacy plumbing. GA4 works with cookies and identifiers, so a European audience usually means a consent banner before it can count anyone — and the banner itself changes what you see.
• The fine print on the data. On a standard GA4 property you can keep user-level and event-level data for 2 months or 14 months, and no longer; the 26-month and 50-month options are reserved for the paid 360 tier.¹ And when a row contains too few users, GA4 withholds it: data thresholds are system-defined, you cannot adjust them, and the only remedies Google documents are widening the date range or exporting to BigQuery.²

None of this is a scandal — it is a design decision made for large properties with legal teams and analysts. My point is narrower: on a site with a few thousand visits a month, you are paying the complexity tax of a tool built for someone else's questions.

The three questions a small site really asks

After twenty years of sitting in front of these dashboards, I would bet that most owners of small sites ask three things and no more:

1. Where did my traffic come from? Which source, which campaign, which page brought the visitor.
2. Which pages bring me business? Not which pages get read — which lead to the enquiry, the order, the booking. That only works with a properly defined goal, and most small sites define it as a click and call it done. Once the goal is right, the metric that ranks those pages is revenue per session: money per visit, not visits.
3. Where do people get stuck? The step, form or page where visits stop becoming customers. This is the question with money attached to it, and it is the one a chart answers least well: a high exit rate tells you where the leak is, never why.

A fourth question arrives as soon as the first three have answers: why did they get stuck? No dashboard answers that one. The answer is not in the numbers but in the sessions behind them — which is why a stack of numbers alone has never changed a page for the better.

GA4 vs the lighter options, compared

Here is the honest landscape for a site with a few thousand visits a month and no data team. Prices move, so I describe the shape of each option instead.

Read the table again and notice what decides the answer: not features, the question. A shop running ads on four channels needs GA4 or a serious equivalent, because attribution is its question. A consultant with a five-page site and a contact form does not — theirs is which page produces enquiries and where the form loses people.

What you give up when you leave GA4

Know this before you uninstall anything: the switching cost is real, and mostly invisible at the start.

• Audiences and remarketing. If you retarget visitors or build audiences for ad platforms, GA4 is the piece of plumbing that feeds them. Leaving means rebuilding that elsewhere or losing it.
• Raw event data. Explorations, BigQuery exports, Looker Studio connectors and the whole ecosystem of things built on top of GA4 stop being available to you.
• Free-form questions. In GA4 you can cross almost any dimension with any other. A lighter tool gives you the answers it was designed to give — usually the point, occasionally a wall.
• Your history. Analytics history does not move. A new tool starts counting on the day you install it, so keep the GA4 property around, read-only if you like, until you genuinely no longer look back.

If two or more of those matter to you, the honest answer is that GA4 is probably your tool and the work to do is making it legible, not replacing it.

What I would set up instead (my opinion)

This is a judgement, not a law, and it comes from twenty years of cleaning up measurement setups that nobody used.

For a site under roughly 50,000 pageviews a month with no data team, I would install one lightweight tool that answers the three questions in a single place, and stop there. Not because specialists are bad, but because every extra script is another thing to keep alive, another consent question, and another dashboard nobody opens. Then, and only when a step is leaking, I would watch the sessions behind it and read the heatmap of that page — without over-interpreting it.

I would keep GA4 in exactly two situations: there is paid media in the mix, or someone outside the business asks for it. Otherwise it becomes the property you log into twice a year and misread once.

That is the gap SessionInsight was built for: analytics, session replays, heatmaps and goals in one lightweight script, with a cookie-less mode for basic analytics. The free plan covers one website, 50,000 pageviews a month, 10 heatmaps and 10 goals over 30 days; Growth is 9 €/month for 15 sites and a year of history. You can see how it works or read the pricing before you decide anything.

How to switch without losing your bearings

Do not uninstall anything on the day you decide. The switch is a month-long comparison, and it goes like this.

1. Write the questions down first. Three lines on paper. If you cannot write them, no tool will help — and the reasons visitors leave stay a mystery.
2. Install the new tool alongside GA4. Two analytics scripts on one page do not interfere with each other. Leave them both running for four to six weeks.
3. Expect the numbers to disagree, and know why. Consent state, blocked scripts, bot filtering and a different definition of a session are enough to explain a gap; GA4's thresholds on small rows widen it. Treat a 10% difference as normal, a tenfold one as a tracking bug to find.
4. Compare decisions, not decimals. Same top three sources? Same pages producing enquiries? Same leaking step? If the answers match, you can make the same decisions in a tool you actually open.
5. Retire what you do not use. Keep the old property read-only rather than deleting it, and remove the extra script from the page so your consent notice matches what actually runs.
6. Then fix one thing at a time. Note the date of the change and compare the trend afterwards. On a low-traffic site, that is the method that works — there is a way to do it without statistical significance.

FAQ

Sources

1. Google Analytics Help, Data retention (checked 23 September 2026).
2. Google Analytics Help, [GA4] About data thresholds (checked 23 September 2026).